FusionAgent：面向人体识别的动态模型选择多模态智能体
FusionAgent: A Multimodal Agent with Dynamic Model Selection for Human Recognition
把每个生物特征专家模型当作可调用工具，用 MLLM 智能体配合指标驱动的强化微调，为每个测试样本动态选择模型组合，并用基于锚点的置信度 Top-k 融合解决分数错位与嵌入异质性问题；这套「动态模型选择 + 置信度感知融合」思路可直接迁移到多模态或多模型集成的视觉识别系统。
原文依据 we propose FusionAgent, a novel agentic framework that leverages a Multimodal Large Language Model (MLLM) to perform dynamic, sample-specific model selection
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论模型融合是在无约束场景下实现鲁棒识别的关键策略，因为不同模型能够提供互补的优势。这一点在全身人体识别中尤为重要：人脸、步态和体型等生物特征线索在不同样本间存在差异，通常通过分数融合进行整合。然而，现有的分数融合策略通常是静态的，无论样本质量或模态可靠性如何，都会对每个测试样本调用全部模型。为克服这些局限，我们提出 FusionAgent，一种新颖的智能体框架，它利用多模态大语言模型（MLLM）执行动态的、针对具体样本的模型选择。每个专家模型都被视为一个工具，通过带有基于指标奖励的强化微调（RFT），智能体学会为每个测试输入自适应地确定最优模型组合。为解决模型分数错位和嵌入异质性问题，我们引入基于锚点的置信度 Top-k（ACT）分数融合，它以置信度最高的模型为锚点，并以置信度感知的方式整合互补预测。在多个全身生物特征基准上的大量实验表明，FusionAgent 显著优于当前最先进方法，同时通过更少的模型调用实现更高的效率，凸显了动态、可解释且鲁棒的模型融合在真实识别系统中的关键作用。项目主页：https://fusionagent.github.io/。
英文原文摘要 1479 字符
Model fusion is a key strategy for robust recognition in unconstrained scenarios, as different models provide complementary strengths. This is especially important for whole-body human recognition, where biometric cues such as face, gait, and body shape vary across samples and are typically integrated via score-fusion. However, existing score-fusion strategies are usually static, invoking all models for every test sample regardless of sample quality or modality reliability. To overcome these limitations, we propose FusionAgent, a novel agentic framework that leverages a Multimodal Large Language Model (MLLM) to perform dynamic, sample-specific model selection. Each expert model is treated as a tool, and through Reinforcement Fine-Tuning (RFT) with a metric-based reward, the agent learns to adaptively determine the optimal model combination for each test input. To address the model score misalignment and embedding heterogeneity, we introduce Anchor-based Confidence Top-k (ACT) score-fusion, which anchors on the most confident model and integrates complementary predictions in a confidence-aware manner. Extensive experiments on multiple whole-body biometric benchmarks demonstrate that FusionAgent significantly outperforms SoTA methods while achieving higher efficiency through fewer model invocations, underscoring the critical role of dynamic, explainable, and robust model fusion in real-world recognition systems. Project page: https://fusionagent.github.io/.